Bayesian multitarget tracking is an inherently nonlinear problem. Even when the state models and sensor noise associated with individual targets and observations is Gaussian, the "true" data likelihood, as f...
详细信息
ISBN:
(纸本)0819449563
Bayesian multitarget tracking is an inherently nonlinear problem. Even when the state models and sensor noise associated with individual targets and observations is Gaussian, the "true" data likelihood, as formulated within the framework of finite-set statistics, is non-Gaussian. Missed detections and false alarms, combined with the fact that targets may enter and leave the scene at random times, complicate matters further. The resulting Bayesian posterior is analytically foreboding, and many conventional estimators are not even defined. We propose an algorithm for generating samples from the posterior based on jump-diffusion processes. When discretized for computer implementation, the jump-diffusion method falls into the general class of Markov chain Monte Carlo methods. The diffusions refine estimates of continuous parameters, such as positions and velocities, whereas the jumps are responsible for major discrete changes, such as adding and removing targets. Jump-diffusion processes have been previously applied to performing automatic target recognition in infrared images and tracking multiple targets using raw narrowband sensor array and high-resolution range profile data. Here, we apply jump-diffusion to the more traditional class of target tracking problems where raw sensor data is preprocessed into reports, but the report-to-target association is unknown. Our formulation maintains the flavor of other recent work employing finite-set statistics, in that no attempts to explicitly associate specific reports with specific targets are needed.
A new ultra-fast (increasing the calculation rate about 9 orders and more) calculation scheme, enabling one to solve small-angle multi-scattering problems for large (size about 1000 scattering lengths and more) object...
详细信息
ISBN:
(纸本)0819458473
A new ultra-fast (increasing the calculation rate about 9 orders and more) calculation scheme, enabling one to solve small-angle multi-scattering problems for large (size about 1000 scattering lengths and more) objects is described.
Aiming at the large bias of LSE (Least Squares Estimation) in estimating MTBF (mean time between failures) under a small sample of data, a Bayesian MTBF estimating method is proposed for NC (numerical control) machine...
详细信息
ISBN:
(纸本)9781467391665
Aiming at the large bias of LSE (Least Squares Estimation) in estimating MTBF (mean time between failures) under a small sample of data, a Bayesian MTBF estimating method is proposed for NC (numerical control) machine tools. To solve difficulty in directly presenting the prior distributions of Weibull parameters, an expert-judgment method which incorporates prior information is developed to indirectly obtain Weibull parameters' prior distributions. Aiming at the problem that analytic solutions to Weibull parameters' posterior distributions and estimators are impossible to obtain, a metropolis algorithm is developed. The iteration procedure of the algorithm is presented;the posterior distribution of each parameter is simulated;and the parameter estimators and MTBF are obtained. Given the actual MTBF as standard value, the proposed method and LSE are applied to the same real case respectively. The results indicate that when sample size n <= 10, relative errors of the proposed method lie between 4.43% and 7.19%, which are smaller than those of LSE. The proposed Bayesian MTBF estimating method is better than LSE and suitable for NC machine tools under small samples.
We study analytically the relaxation eigenmodes of a simple Monte Carlo algorithm, corresponding to a particle in a box which moves by uniform random jumps. Moves outside of the box are rejected. At long times, the sy...
详细信息
We study analytically the relaxation eigenmodes of a simple Monte Carlo algorithm, corresponding to a particle in a box which moves by uniform random jumps. Moves outside of the box are rejected. At long times, the system approaches the equilibrium probability density, which is uniform inside the box. We show that the relaxation towards this equilibrium is unusual: for a jump length comparable to the size of the box, the number of relaxation eigenmodes can be surprisingly small, one or two. We provide a complete analytic description of the transition between these two regimes. When only a single relaxation eigenmode is present, a suitable choice of the symmetry of the initial conditions gives a localizing decay to equilibrium. In this case, the deviation from equilibrium concentrates at the edges of the box where the rejection probability is maximal. Finally, in addition to the relaxation analysis of the Master equation, we also describe the full eigen-spectrum of the Master equation including its sub-leading eigen-modes.
Many complex problems like Speech Recognition, Bioinformatics, Climatology, Control and Communication are solved using Hidden Markov Models (HMM). Mostly, optimization problems are modeled as HMM learning problem in w...
详细信息
ISBN:
(纸本)9781424429271
Many complex problems like Speech Recognition, Bioinformatics, Climatology, Control and Communication are solved using Hidden Markov Models (HMM). Mostly, optimization problems are modeled as HMM learning problem in which HMM parameters are either maximized or minimized. In general, Baum-Welch Method (BW) is used to solve HMM learning problem giving only local maxima/minima in exponential time. In this paper, we have modeled HMM learning problem as a discrete optimization problem such that randomized search methods can be used to solve the learning problem. We have implemented metropolis algorithm (MA) and Simulated Annealing algorithm (SAA) to solve the discretized HMM learning problem. A comparative study of randomized algorithms with the Baum Welch method to estimate the HMM learning parameters has been made. The metropolis algorithm is found to reach maxima in minimum number of transactions as compared to the Baum-Welch and Simulated Annealing algorithms.
The paper provides a Bayes analysis, based on free-knot spline technique, of the popular autoregressive model having functional-coefficients. The model was initially proposed by Chen and Tsay (1993). The technique of ...
详细信息
The paper provides a Bayes analysis, based on free-knot spline technique, of the popular autoregressive model having functional-coefficients. The model was initially proposed by Chen and Tsay (1993). The technique of polynomial splines of different orders is used to approximate the functional-coefficients. A sample based approach using the Gibbs sampler algorithm with intermediate metropolis steps is adopted to draw the posterior estimates for the parameters involved. Additionally, the technique of reversible jump Markov chain Monte Carlo is incorporated to update the location and number of knots in the polynomial spline. The paper then proceeds with the motive of obtaining both retrospective and prospective predictions based on the selected model. The complete procedure is illustrated by both simulated and a real dataset representing the exchange rate of Indian rupees relative to the US dollars.
In this paper, optimisation algorithms are successfully applied to a number of antenna array beampattern synthesis problems including, a Uniform Linear Array (ULA), a ULA with an inactive element, and a 2D Wireless Se...
详细信息
ISBN:
(纸本)9781467369749
In this paper, optimisation algorithms are successfully applied to a number of antenna array beampattern synthesis problems including, a Uniform Linear Array (ULA), a ULA with an inactive element, and a 2D Wireless Sensor Network (WSN) array. The algorithms presented include a Greedy algorithm (GDA), a metropolis algorithm (MA), and a Genetic algorithm (GA).
This paper provides a general procedure to estimate structural vector autoregressions. The algorithm can be used in constant or time-varying coefficient models, and in the latter case, the law of motion of the coeffic...
详细信息
This paper provides a general procedure to estimate structural vector autoregressions. The algorithm can be used in constant or time-varying coefficient models, and in the latter case, the law of motion of the coefficients can be linear or nonlinear. It can deal in a unified way with just-identified (recursive or nonrecursive) or overidentified systems where identification restrictions are of linear or of nonlinear form. We study the transmission of monetary policy shocks in models with time-varying and time-invariant parameters.
Large scale Potts model Monte Carlo simulation was carried on 3-dimensional square lattices of 100 3 and 200 3 sizes using the metropolis algorithm to study grain growth behavior. Simulations were carried out to inves...
详细信息
ISBN:
(纸本)9783038351795
Large scale Potts model Monte Carlo simulation was carried on 3-dimensional square lattices of 100 3 and 200 3 sizes using the metropolis algorithm to study grain growth behavior. Simulations were carried out to investigate both growth kinetics as well as the Zener limit in two-phase polycrystals inhibited in growth by second phase particles of single-voxel size. Initially the matrices were run to 10,000 Monte Carlo steps (MCS) to check the growth kinetics in both single phase and two-phase poly-crystals. Grain growth exponent values obtained as a result have shown to be highest (similar to 0.4) for mono-phase materials while the value decreases with addition of second phase particles. Subsequently the matrices were run to stagnation in the presence of second phase particles of volume fractions ranging from 0.001 to 0.1. Results obtained have shown a cube root dependence of the limiting grain size over the particle volume fraction thus reinforcing earlier 3D simulation efforts. It was observed that there was not much difference in the values of either growth kinetics or the Zener limit between 100(3) and 200(3) sized matrices, although the results improved mildly with size.
In this paper, optimized communication interfaces in which users select phonemes (sounds) instead of letters or whole words are presented and evaluated. Optimization is based on phoneme transition likelihoods (i.e., t...
详细信息
ISBN:
(纸本)9781450349260
In this paper, optimized communication interfaces in which users select phonemes (sounds) instead of letters or whole words are presented and evaluated. Optimization is based on phoneme transition likelihoods (i.e., the probability of transitioning from one phoneme to another in a particular communication corpus), similar to letter-to-letter transition likelihoods used to optimize orthographic interfaces. However, it is unknown to what extent phoneme transition likelihoods vary by corpus, nor how optimizing based on different corpora affects the final interface efficiency. Here we used computational evaluations to compare phoneme transition likelihoods between various phonemic corpora and optimize phonemic interfaces with each corpus. Each interface's efficiency was evaluated against all the corpora. Phoneme-to-phoneme transitions were highly correlated across corpora (r = 0.7-0.86). Optimization based on phoneme-to phoneme transition likelihoods improved efficiency by around 2030% compared to random phonemic layouts, regardless of the corpus used to optimize the interface. Optimizations using different corpora were similar, varying only by 3-5%. We conclude that, if possible, future phonemic interfaces should be optimized via a corpus from the intended user's communication. If this is not possible, however, optimization still improved efficiency using all testing corpora, suggesting that optimizing via any relevant corpus is indicated over other layouts.
暂无评论